ValueError: Checkpoint was expecting to be a trackable object (an object derived from Trackable)

Issue type

Bug

TensorFlow version

tf 2.10.0

Custom code

Yes

OS platform and distribution

Windows 10 Enterprise

Python version

3.9.16

Current behavior?
I'm receiving an error when I try to restore the model checkpoint. I've seen a posting on here that's similar, but I think my case is different. Help is very much appreciated!

I'm using a pre-trained object detection model called SSD MobileNet V2 FPNLite 320x320

Standalone code to reproduce the issue

 import os
import tensorflow as tf
import pandas as pd
import openpyxl
import cv2 
import numpy as np

from object_detection.utils import label_map_util
from object_detection.utils import visualization_utils as viz_utils
from object_detection.builders import model_builder
from object_detection.utils import config_util
from matplotlib import pyplot as plt
from pathlib import Path


os.chdir(r"C:\Users\mill286")

CUSTOM_MODEL_NAME = 'my_ssd_resnet50_v1_fpn' # *** Enter here the name of the model you trained. ***
files = {
    'PIPELINE_CONFIG':os.path.join('tensorflow', 'workspace','models', CUSTOM_MODEL_NAME, 'pipeline.config')
}
# Load pipeline config and build a detection model
configs = config_util.get_configs_from_pipeline_file(files['PIPELINE_CONFIG'])
detection_model = model_builder.build(model_config=configs['model'], is_training=False)
# Restore checkpoint
ckpt = tf.compat.v2.train.Checkpoint(model=detection_model)
ckpt.restore(os.path.join(paths['CHECKPOINT_PATH'], 'ckpt-54.index')).expect_partial() # *** Replace the number in 'ckpt-XX' with the checkpoint you want to use. *** 

Relevant log output

ValueError                                Traceback (most recent call last)
Cell In[18], line 2
      1 # Restore checkpoint
----> 2 ckpt = tf.compat.v2.train.Checkpoint(model=detection_model)
      3 ckpt.restore(os.path.join(paths['CHECKPOINT_PATH'], 'ckpt-54.index')).expect_partial()

File ~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\checkpoint\checkpoint.py:2142, in Checkpoint.__init__(self, root, **kwargs)
   2140 if isinstance(converted_v, weakref.ref):
   2141   converted_v = converted_v()
-> 2142 _assert_trackable(converted_v, k)
   2144 if root:
   2145   # Make sure that root doesn't already have dependencies with these names
   2146   child = trackable_root._lookup_dependency(k)

File ~\Anaconda3\envs\tensorflow\lib\site-packages\tensorflow\python\checkpoint\checkpoint.py:1562, in _assert_trackable(obj, name)
   1559 def _assert_trackable(obj, name):
   1560   if not isinstance(
   1561       obj, (base.Trackable, def_function.Function)):
-> 1562     raise ValueError(
   1563         f"`Checkpoint` was expecting {name} to be a trackable object (an "
   1564         f"object derived from `Trackable`), got {obj}. If you believe this "
   1565         "object should be trackable (i.e. it is part of the "
   1566         "TensorFlow Python API and manages state), please open an issue.")

ValueError: `Checkpoint` was expecting model to be a trackable object (an object derived from `Trackable`), got <object_detection.meta_architectures.ssd_meta_arch.SSDMetaArch object at 0x000001E163D93910>. If you believe this object should be trackable (i.e. it is part of the TensorFlow Python API and manages state), please open an issue.